A prospective follow-up study of chest pain patients - with emphasis on patients with panic disorder.
Christine Bull Bringager
Department of Psychiatry and
Department of Cardiology Ullevål University Hospital
Faculty of Medicine University of Oslo
Oslo 2007
© Christine Bull Bringager, 2007
Series of dissertations submitted to the Faculty of Medicine, University of Oslo No. 550
ISBN 978-82-8072-250-8
All rights reserved. No part of this publication may be
reproduced or transmitted, in any form or by any means, without permission.
Cover: Inger Sandved Anfinsen.
Printed in Norway: AiT e-dit AS, Oslo, 2007.
Produced in co-operation with Unipub AS.
The thesis is produced by Unipub AS merely in connection with the thesis defence. Kindly direct all inquiries regarding the thesis to the copyright holder or the unit which grants the doctorate.
Unipub AS is owned by
The University Foundation for Student Life (SiO)
TABLE OF CONTENTS
Preface……… 5
List of papers……… 5
Acknowledgements ……….... 6
Abbreviations……… 8
General introduction………... 9
1.1.0. Chest pain………... 9
1.1.1. Chest pain- epidemiology and etiology……….. 9
1.1.2. Chest pain prognosis………... 10
1.2.0. Panic disorder………. 10
1.2.1. Panic disorder- prevalence……….. 11
1.2.2. Panic disorder and chest pain……….. 11
1.2.3. Outcome of panic disorder in chest pain patients………... 12
1.2.4. Outcome of panic disorder in clinical psychiatric settings………. 12
1.2.5. Predictors of panic disorder outcome………. 15
1.2.6. Treatment of panic disorder……… 15
1.2.7. Panic disorder and cardiac morbidity and mortality………... 16
1.3.0. Non-fearful panic disorder………. 17
Aims of the study………. 18
2.1. Prevalence of panic disorder at follow-up………. 18
2.2. Outcome of baseline panic disorder at follow-up……….. 18
2.3. Identify predictors of continued PD………... 18
2.4. Treatment and perceived treatment need……… 18
2.5. Effect of PD on chest pain, CAD and mortality………. 18
2.6. Prevalence and outcome of nonfearful panic disorder in chest pain patients………… 19
Material and methods... 20
3.1.0. Patient sample………. 20
3.1.1. Baseline………... 20
3.1.2. One year follow-up………. 20
3.1.3. Nine year follow-up……… 21
3.2.0. Methods……….. 21
3.2.1. Assessments of sampling bias……….... 21
3.2.2. Procedure……… 22
3.2.3. Psychiatric assessments……….. 23
3.2.4. Cardiological assessments………... 27
3.2.5. Statistical analysis……….. 28
3.2.6. Ethical aspects……….... 29
Summary of papers………. 30
General discussion……….. 34
5.1.0. Methodological issues……… 34
5.1.1. Sample……… 34
5.1.2. Assessments……….... 36
5.1.2.1. Psychiatric assessments………... 36
5.1.2.1. Cardiological assessments……… 39
5.2.0. Results and clinical implications………... 40
5.2.1. Prevalence of panic disorder at follow-up……….. 40
5.2.2. Outcome of baseline panic disorder at follow-up………... 41
5.2.3. Predictors of persistent PD………. 43
5.2.4. Treatment and perceived treatment need……….... 45
5.2.5. Effect of PD on chest pain, CAD and mortality……….. 47
5.2.6. Prevalence and outcome of nonfearful panic disorder (NFPD)……….. 48
General conclusions……….... 51
Suggestions for future research ……….... 53 Tables……….. 54
References... 56
Preface
The present thesis is based on the papers listed below, referred to in the text by their Roman numbers in brackets.
List of papers:
I. Dammen T, Bringager CB, Arnesen H, Ekeberg O, Friis S.” A 1-year follow-up study of chest-pain patients with and without panic disorder”. Gen Hosp Psychiatry 2006; 28(6):516-524.
II. Bringager CB, Friis S, Arnesen H, Dammen T. “Nine year follow-up of panic disorder in chest pain patients: Clinical course and predictors of outcome”.
Psychosomatic Medicine. Submitted.
III. Bringager CB, Husebye T, Friis S, Arnesen H, Dammen T. “A long-term follow-up study of chest pain patients: Effect of panic disorder on mortality, morbidity and quality of life”. Cardiology. In Press.
IV. Bringager CB, Dammen T, Friis S.”Nonfearful panic disorder in chest pain patients”. Psychosomatics 2004; 45:69-79
V. Bringager CB, Gauer K, Arnesen H, Friis S, Dammen T. ”Nonfearful panic disorder in chest pain patients: Status after nine-year follow-up”. Psychosomatics.
In press.
Acknowledgements
Many people have contributed to the completion of this thesis.
The study was conducted as collaboration between the Department for Research and Education, Psychiatric Division and Cardiology Outpatient Clinic at Ullevål University hospital.
First, I am deeply indebted to my outstanding supervisor team consisting of Associate Professor Toril Dammen, Professor Svein Friis, and Professor Harald Arnesen, for their enthusiasm and encouragement throughout the research process. The work could not have been carried out without them! Toril Dammen has been my main supervisor and introduced me to her previous study of chest pain patients and generously gave me access to her baseline and one year follow-up data. She has guided me through the research process from the first idea that a long-term follow-up study should be undertaken to the final thesis has been completed.
Svein Friis has step-by-step taught me a scientific way of thinking. He has willingly shared his statistical knowledge and always been available for me when I struggled with the data. Harald Arnesen has been the coordinator of the cardiology investigations and managed to fit in 150 additional patients in the routine schedule of a very busy cardiology outpatient clinic. He has consciously read through my sometimes long and wordy papers and enthusiastically discussed psychiatric issues.
I owe great appreciation to the staff at the Cardiology Outpatient clinic. Trygve Husebye for evaluating the diagnostic considerations made at the bicycle exercise test, Christin von der Lippe for scheduling the study patients, and the doctors, nurses and office staff involved in the planning and execution of the exercise tests.
All the patients who participated in the follow-up assessments need a special appreciation.
They generously spent their time taking part in the exercise test, interview and filled in self- report questionnaires.
I am also grateful to Leiv Sandvik who has regularly given me statistical supervision and at the same time shared some of his philosophy of life. I want to thank Katrine Gauer who showed up during the winter of 2005 with a special interest in non-fearful panic disorder and worked with
me on paper five. I thank Thomas Small for managing the liaison services without me during times when I was busy collecting data or otherwise needed time to focus on the study. I thank Bjørg Riise for secretarial help and Oddmar Moen for assistance when the computer caused me trouble. I also want to thank all the participants of our research meetings who have given me useful criticism and encouragement and I thank my colleague Jan Ivar for cheering me up when I thought my mission was impossible.
And last but not least, I am indebted to my family and friends; my mother who has taught me to believe in my self and always supported my choices, my sister Benedicte who has been a role model in the academic world, my mother-in-law Siri for being a frequent baby sitter, and my friends for not forgetting me.
And finally I will thank my daring husband Nils and my beautiful children Fredrik and Helene for their love and support.
Abbreviations
ACQ Agoraphobic Cognitions Questionnaire ANOVA Analysis of variance
B Beta
BMI Body mass index CAD Coronary artery disease
CABG Coronary artery by-pass grafting CBB Christine Bull Bringager CI Confidence interval
DSM-IV Diagnostic and Statistical manual of Mental Disorders, 4th edition DSM-III-R Diagnostic and Statistical manual of Mental Disorders, 3rd edition, revised D(u) Observed number of deaths in the study sample
ECG Electrocardiography ED Emergency department EPQ Eysenck Personality Questionnaire GAD Generalized anxiety disorder
HARP Harvard-Brown Anxiety Research Project HRV Heart rate variability
IAS Illness Attitude Scale HRQOL Health related quality of life MIA Mobility Inventory for Agoraphobia
MOS SF-36 The Medical Outcomes Study Short Form-36 NCCP Non-cardiac chest pain
NFPD Nonfearful panic disorder NoPD No panic disorder Non-PD No panic disorder
OR Odds ratio
PD Panic disorder
PDQ Personality Diagnostic Questionnaire PCI Percutanous coronary intervention RR Relative risk
SCID-I Structured Clinical Interview for DSM-IV- Axis I disorders SCL-90-R Symptom Checklist-90-Revised
SD Standard deviation
SF MPQ Short Form McGills Pain Questionnaire SMR Standardized mortality ratio
SPSS Statistical Package for Social Sciences TAS-20 The 20-itemToronto Alexithymia Scale
TD Toril Dammen
VAS Visual Analogue Scale
Introduction
1.1.0 Chest pain
1.1.1. Chest pain epidemiology and etiology
Chest pain is a common symptom in the general population. According to a population- based study, the lifetime prevalence of chest pain is approximately 39% and 33% have non- cardiac chest pain (NCCP) (i.e. not diagnosed as ischemic heart disease or not classified as angina according to the Rose Angina Questionnaire criteria: substernal pain precipitated by exertion and relieved by rest) (1). About 23% of patients with non-cardiac chest pain visit a doctor because of this symptom and male sex, increasing age, mild symptoms and anxiety are associated with health care seeking (1). Moreover, 12% suffer from chronic chest pain which is the most common reason for referral to cardiac outpatient clinics (2;3).
In about 50% of chest pain patients referred to cardiology clinics there is no evidence of coronary artery disease (4). Still, chest pain patients often go through costly and time-
consuming investigations that put an economic burden on the health care system (5). Heart related disorders like micro vascular ischemia and mitral valve prolapse are suggested to be causes of chest pain in a minority of patients, despite normal angiograms (6). Gastro- esophageal disorders, especially esophageal reflux, are thought to explain the symptoms in about 30% of cases. Additionally, musculoskeletal disorders, pulmonary disorders, other gastrointestinal disorders and psychiatric disorders (major depression, anxiety disorders and somatoform disorders) are also reported as non-cardiac causes of chest pain (3;5-10).
Panic disorder (PD) is the anxiety disorder most frequently associated with chest pain as it occurs in 10 to 60 % of chest pain patients with or without coronary artery disease in cardiology clinics or emergency departments (11-14). Cross sectional studies of PD in chest pain patients have reported associations with high psychological distress, high rates of co- morbid psychiatric disorders, suicidal ideation, work disability, quality of life impairment and high medical utilization (12-14). However, there is a lack of longitudinal studies of chest pain patients with PD. To the best of our knowledge there are only two previous studies (15;16) which both have major shortcomings as described in section 1.2.3. One of the main purposes of the present study was therefore to explore the longitudinal course of PD and the effect of PD on the short and long-term outcome of chest pain patients (papers I-III).
1.1.2. Prognosis of chest pain patients.
The results of studies on mortality in chest pain patients are somewhat inconsistent. Most studies investigating the relationship between NCCP and cardiac morbidity and mortality, are conducted with NCCP patients with negative coronary angiography. These studies have reported an excellent mortality prognosis with survival rates of 91-98% up to 12 years after the angiographic assessments (17-19). In contrast, two population-based studies reported that the mortality rate of chest pain patients is higher than that of the normal population. One of these studies concerned men with “possible angina”, which was detected using the World Health Organization’s Angina Questionnaire, and normal exercise test (20). The other concerned men with non-specific chest pain that was not considered typical of angina (21). Both studies reported a significantly greater incidence of cardiovascular mortality in patients with chest pain than in those with no symptoms of angina after 26 and 16 years respectively (RR 1.97-2.46).
Despite the excellent prognosis regarding mortality and risk of developing cardiac disorder in patients with negative angiography, as much as 70% of the patients still have regular chest pain at follow up 1-11 years after cardiac investigation (18;22-24) and about 50%
continue to believe they have a heart condition (24;25). Many (71-79%) are continuously treated as having a heart condition by their physicians by prescribing heart medication (18;25).
Additionally, NCCP patients report physical and social disability and it has been reported that 50% are unable to work at follow-up (24).
The reason why these patients have such a poor prognosis is largely unexplored. Despite the high prevalence of PD in chest pain patients, there is a scarcity of research addressing the association between the presence of PD and the long-term outcome of these patients.
1.2.0. Panic disorder
Panic disorder is characterized by recurrent panic attacks followed by anticipatory anxiety or avoidance behavior which interfere with the patients work, family and social life.
The panic attacks are unexpected and consists of at lest four of 13 somatic or cognitive symptoms: palpitations, sweating, trembling, shortness of breath, choking, chest pain, nausea, dizziness, derealization or depersonalization, paresthesias, chills or flushes, fear of dying, fear of going crazy or losing control. The DSM IV criteria for PD are shown in table 1. There is evidence that the age of onset of PD is bimodal (26). The majority have their first panic attack in the twentieths (27) but there is also a group with age of onset between age 50 and 60 (late- onset panic disorder) (26;28;29). It is well documented that PD is associated with physical and
social disability, occupational dysfunction, quality of life impairment and health care utilization in psychiatric and primary care patients (30-33).
1.2.1. Prevalence of panic disorder
According to the US National Co morbidity Survey Replication, the lifetime prevalence of PD in the general population is estimated to 4.7% and the 12 month prevalence to 2.8 % (27). Furthermore, 3 to 8 % of patients seen by primary care physicians suffer from PD (34).
The rate of PD is considerably higher in patients who present with symptoms unexplained by a medical disorder. Studies have reported the following prevalence’s of PD in patients with various conditions; irritable bowel syndrome (29-38%), medically unexplained dizziness (13%) and chronic fatigue (13-30%) (35). Among chest pain patients the prevalence of PD is estimated to 25-60% among those with NCCP (11;36) and 10-50% among those with coronary artery disease (CAD) (37). The rather large divergence in prevalence is due to differences in settings (i.e. family practice, emergency or cardiology departments) and chest pain
characteristics (i.e. typical vs. atypical chest pain).
1.2.2. Panic disorder and chest pain.
Chest pain is one of the symptoms included in the list of diagnostic criteria for panic attacks and 22-70% of PD patients experience chest pain as one of the symptoms of an attack (38). PD patients often relate their chest pain to heart disease (39) even after CAD is ruled out by angiography (15). The fear of having heart disease causes frequent referral to emergency departments or cardiology clinics with costly cardiological investigations as the result (11;33;35).
Efforts have been made to discriminate between chest pain associated with either PD or CAD. In PD patients the chest pain is more often atypical as opposed to CAD patients who have more typical chest pain (defined as substernal, exertional and relieved by nitroglycerine) (11). However, PD patients may also have typical chest pain and panic attacks may in fact cause cardiac ischemia in patients with coronary artery disease (40). In a considerable proportion of chest pain patients PD and CAD co-exist (11;37). Therefore, it is difficult to exclude one diagnosis or the other just by clinical judgment. Subsequently Katerndahl proposed that when either CAD or PD is recognized by a primary care physician, the other should also be considered (11).
1.2.3. Outcome of panic disorder in chest pain patients.
The outcome of panic disorder in chest pain patients have been reported in two previous follow-up studies. Beitman and colleagues (15) investigated the outcome after three years of 72 patients with chest pain and negative coronary angiography of whom 36 suffered from PD.
They found an association between PD at baseline and prevalence of chest-pain episodes at follow-up. PD patients were more convinced that their symptoms were heart-related and rated themselves as more disabled at follow-up compared to the patients without PD. Fleet and colleagues (16) conducted a two-year follow-up of chest pain patients admitted to an emergency department and reported higher chest pain prevalence and medical utilization among the patients with PD. Furthermore, the PD patients were more impaired at follow-up regarding panic-anxiety symptoms, agoraphobic avoidance, and reported more frequently poor general health and suicidal ideation (32%).
However, these studies are limited by small sample size (n=72) (15), low-participation rate at follow-up (54%) (16) and lack of diagnostic reassessments (15;16) regarding PD and CAD. Therefore the outcome of PD in patients presenting with chest pain presenting in cardiology clinics is largely unexplored.
1.2.4. Outcome of panic disorder in psychiatric settings.
Course of panic disorder
A methodological challenge with research addressing the course of PD is that there is no consensus with regard to criteria for remission and relapse. In 1994 a conference report was published making recommendations on standardized assessments for panic disorder research (41). However, the group was unable to come to an agreement regarding definitions of
“remission”, “recovery” and “relapse”. Therefore, studies of PD have used different criteria for remission which make comparisons difficult.
In the Harvard Brown Anxiety Research Project (HARP) (42), the intensity of panic attacks and agoraphobia was graded on a rating scale from one to six where one indicated a symptom free state and six indicated daily panic attacks or extensive phobic avoidance. Full remittance of PD was defined as eight consecutive weeks of symptom ratings “one” or “two”
for both panic and agoraphobia. Relapse was defined as one week of rating “five” or more after a period of remittance. The HARP study included PD patients (n=309) admitted to one of 11 anxiety disorder clinics in the Massachusetts area who were followed monthly for two years.
They reported a chronic course of PD with high risk of relapse after remission of PD symptoms (42). After 12 years, the probability of remission was 82% for patients with PD without
agoraphobia and 48% for patients with PD with agoraphobia, but the probability of recurrence after a period of remission was more than 50% for both groups (43).
Katschnig and colleagues (44) included PD patients (n=1647) who were recruited from psychiatric clinics and through news media to participate in clinical drug trials (The Cross- National Collaborative Panic Study Phase I or II). They did not use criteria for remission and relapse of PD, but reported the panic attack frequency during the week and month preceding the follow-up interview by using the Panic Attack Scale. In addition, global phobia was rated on an 11-point scale. Four years after the end of the trials 25% of the patients were reassessed (45) and the authors reported subgroups of PD courses where 31% remitted and stayed well, 24% had a fluctuating course and 45% had a more chronic debilitating course. Furthermore, cohorts of participants of this study have been reassessed after six, 11 and 15 years (46-48).
After six years and 11 years 75% and 87.5% reported no panic attacks while after 15 years 18% still met criteria for current PD according to DSM III.
The results of follow-up studies of PD patients referred to psychiatric treatment may not be generalized to PD patients seen in other medical settings as there is some evidence that few PD patients in medical settings receive effective anti-panic medication or psychotherapy (15;49;50). Because the follow-up studies of PD in chest pain patients have not made diagnostic reassessments, the stability of PD among these patients is unexplored.
Outcome of panic disorder.
Symptoms of anxiety and phobic avoidance. Although many patients do not fulfill criteria for PD at follow-up and the majority of patients improve regarding symptoms of anxiety and phobic avoidance, few are symptom-free (44;51). Moreover, residual anxiety and avoidance symptoms are of negative prognostic significance for panic relapse (52). The knowledge regarding symptom reduction/persistence in PD is derived from studies of patients who have received antipanic medication or psychotherapy. In addition, a substantial proportion of patients still use medication at the time of follow-up, therefore the true natural course of symptoms is unknown.
Psychiatric co-morbidity. One cross-sectional study (53) reported that 61% of patients referred for treatment of PD suffer from at least one other anxiety or affective disorder, but the overall co-morbidity decreased when PD was treated. A recent epidemiological study (27) reported that 83% of PD patients without agoraphobia and 100% of PD patients with
agoraphobia suffered from at least one other psychiatric disorder and agoraphobia was found in about one fourth of patients with lifetime PD. Co-existence of agoraphobia and PD is far
higher in patients seen in psychiatric clinics (75-80%) (32;44;54). One study reported that agoraphobia is less readily treated than panic attacks (55) but still tend to decrease over time (44;46). Agoraphobia is most commonly viewed as a consequence of panic attacks, but it may also precede the onset of attacks (52). The rate of co-morbid generalized anxiety disorder (GAD) is also significant and in some cases GAD may rather be secondary to panic symptomatology (53;54). In those cases, GAD symptoms may diminish when PD is treated (53). Otherwise, GAD generally has a more chronic and persistent course than PD (43).
Regarding major depression, about 60% of PD patients experience one or more depressive episode during the lifetime course of their illness while about 30% of PD patients have experienced a depressive episode during the last year (56). The rate of major depression and alcoholism tend to increase over time in some follow-up studies of PD (47), yet, not in all (46).
The change in co-morbidity between PD and other psychiatric disorders over time has not been addressed in previous studies of PD in chest pain patients.
Suicidal ideations and attempts. Recent research have revealed panic disorder as an independent risk factor for suicidal ideation in primary care patients (57); and a prospective population based study reported anxiety disorders, and among them PD, as risk factors for future suicidal ideation and attempts (58) when co-morbid disorders were controlled for. There is also some evidence that suicidal ideation increases over time in chest pain patients with PD (16), but the long-term association between PD and suicidal ideation in chest pain patients is unknown.
Quality of life. The association between PD and quality of life impairment has been established in several cross-sectional studies of PD patients in community samples (59;60) and in psychiatric settings (31;61). Health related quality of life (HRQOL) is one of the targets for treatment of PD (62), but with some exceptions, the longitudinal outcome of PD patients with regard to HRQOL has hardly been addressed. Previous studies have used the Sheehan Disability Scale to assess quality of life of PD patients. Katschnig et al.(63) reported a significant improvement in disability of three life domains. After four years the proportions of patients reporting no or mild disability was 82% regarding work, 77% regarding family life and 70 % regarding social life while the proportions at baseline were 22%, 19% and 7%
respectively. Carpinello et al. (32), came to the opposite conclusion as they reported significant disability in 60% of their PD patients and 40% were dissatisfied with at least 50% of life domains. These discrepancies may be explained by the naturalistic design of the studies.
Furthermore, two studies have used HRQOL measurements to evaluate treatment outcome of PD. They have reported significant short-term (six weeks and 16 weeks) improvement in
quality of life measured by the MOS Short Form 36 in clonazepam (64) and imipramin (65) treated patients. However, no previous research has addressed the outcome of PD in chest pain patients in terms of HRQOL impairment.
1.2.5. Predictors of panic disorder outcome.
The majority of studies that have attempted to establish predictors of PD outcome have focused on factors that affect PD remission or relapse of PD symptoms after a period of remission. The results have been somewhat inconsistent which may be explained by the use of different remission criteria and varying lengths of follow-up (two to eight years). Co-morbid agoraphobia has in most follow-up studies of PD predicted decreased likelihood of PD remission (43;66;67). One study reported that co-morbid major depression decreased the likelihood of remission from PD with agoraphobia but not of PD without agoraphobia, but it increased the odds of relapse of both types of PD (43). Other studies have not found an association between major depression and time to PD remission (66;67). Moreover, presence of personality disorders decreased the probability of achieving a period of PD remission in one study (67) but not in another (68). Co-morbid anxiety disorders have also been associated with a decreased remission rate (67). Regarding demographic variables, women tend to be more likely to remit than men, but their relapse rate was three-fold higher (69). Low socioeconomic status has also predicted longer time to remission from PD (66). Furthermore, in one study disability at baseline was the only variable that predicted disability at four year follow-up (63).
These studies have been conducted in PD patients in clinical or research settings and no previous study of chest pain patients with PD have focused on predictors of PD outcome. Such knowledge is of crucial importance when deciding which patients are in greatest need of treatment.
1.2.6. Treatment of PD
In a recent meta-analysis of PD treatment, a combination of psychotherapy (21 of 23 studies used behavioral or cognitive behavioral therapy) and anti-depressants or psychotherapy alone were recommended as first choice of treatment (70). According to recommendations based on proceedings from the World Council of Anxiety meeting in 2001(71), the first choice of medical treatment for PD is selective serotonin reuptake inhibitors, but high-potency benzodiazepines, reversible monoamine oxidase inhibitors and tricyclic antidepressants are also efficient. The recommended treatment length is 12 to 24 months and for some patients the treatment should be life-long (71;72).
Despite specific and well-documented treatment recommendations, a recent
epidemiological study of PD reported that only about 40% of PD patients receive treatment meeting published treatment guidelines (27). Among PD patients in primary care or cardiology settings the rate of treatment of PD is even lower. Cross-sectional studies have reported that 28% of PD patients in primary care received pharmacological antipanic treatment and 12%
received cognitive therapy (49;50), while amongst chest-pain patients with PD less than 10%
received antipanic treatment (14). A previously reported hypothesis regarding low rates of PD treatment in chest pain patients, is that the PD patients are often unrecognized (13). However, in two follow-up studies of PD in chest pain patients, only 22% (16) and 33% (15) received some kind of PD treatment even though PD was identified. Because treatment may affect PD outcome it is necessary to record it in follow-up studies.
1.2.7. Panic disorder and cardiac morbidity and mortality.
Previous studies have suggested that there is an association between PD, cardiovascular mortality, and CAD (73;74). In the 1980’s, Coryell and colleagues (75;76) reported an increased risk of cardiovascular deaths among inpatients and outpatients with ‘anxiety neurosis’. However, the number of deaths in these studies were low (six and four respectively) and systematic psychiatric diagnostic assessments and causes of death were lacking.
Furthermore, some epidemiological studies have reported a significant association between sudden cardiac death and self-reported panic-like symptoms in healthy men and women (77- 81). One hypothesis attributes a possible link between PD and CAD to an elevated incidence of risk factors for CAD in PD patients (11;82). Another hypothesis emphasizes decreased heart rate variability in PD patients which increases the risk of ventricular arrhythmias and sudden cardiac death (83). However, an extensive review published in 1998 concluded that studies of associations between PD, CAD, and cardiovascular mortality suffer from serious
methodological limitations, such as the lack of sound methodological assessment of PD and prospective design (84). Since then, a recently published cohort study of nearly 40.000 persons diagnosed with PD reported that having a PD diagnosis almost doubled the risk for subsequent CAD (85).
Yet, the association between PD and cardiac morbidity and mortality is controversial and this association has not previously been addressed in prospective follow-up studies of chest pain patients.
1.3.0. Nonfearful panic disorder
Nonfearful panic disorder (NFPD) is a type of PD that was first described by Beitman and colleagues (86) in 1987 among cardiology patients who had panic attacks without the experience of fear. These patients met the DSM III-R criteria for PD by reporting attacks of intense discomfort and at least four of the twelve remaining symptoms on the screening checklist, but they did not report subjective free floating anxiety or fear of dying, fear of ‘going crazy’, or doing something uncontrolled.
Three previous studies in cardiology and emergency department settings have estimated the prevalence of NFPD to 32–44% among PD patients (86-88). However, the concept of NFPD is somewhat controversial. It is not a DSM-IV illness and it remains equivocal whether NFPD is a subgroup of PD or whether it is a distinct diagnostic entity (88;89). Questions have also been raised as to whether NFPD may be better classified as a somatoform disorder, or whether NFPD patients have an undetected medical disorder that is wrongly identified as NFPD (88). Moreover, alexithymia (not having words to express feelings) has been proposed as a possible explanation for NFPD symptoms (90).
To establish NFPD as a well-defined panic disorder subtype would aid its accurate recognition and make an incentive for treatment. So far, it seems reasonable to regard NFPD as a variety of PD based upon cross-sectional comparisons including one longitudinal study suggesting that NFPD is highly similar to PD with fear (86-88;91). One study has found that NFPD patients respond similarly to PD patients to lactate infusions and treatment with anxiolytic medication (92).
Still, more research is required to confirm the validity of NFPD as a subtype of PD. The results of previous studies of NFPD need to be replicated in other PD samples and data regarding the predictive validity of NFPD are currently inadequate (i.e. regarding long-term course). One previous study has addressed the two-year course of NFPD but lacks diagnostic reassessment of PD and NFPD(88).
Aims of the study
The primary aim of the study was to investigate the prognosis of chest pain patients when they are encountered in a cardiology outpatient clinic with special emphasis on the importance of panic disorder:
2.1. Prevalence of panic disorder at follow-up (Papers I and II).
(1) Estimate the prevalence of PD after one year and after nine years among chest pain patients diagnosed with PD at the first examination.
2.2. Outcome of baseline panic disorder at follow-up (Papers I-III).
(1) Compare the patients with and without PD at baseline regarding co-morbid psychiatric disorder, psychological distress, panic-anxiety symptoms, suicidal ideation, hypochondriacal features and health related quality of life after one year and after nine years.
(2) Compare the patients with and without panic disorder at baseline regarding change in scores of the outcome measures outlined above from the first examination to follow-up after one and after nine years.
(3) Compare patients with persistent PD after nine years to those with PD in remission regarding outcome measures listed above.
2.3. Predictors of having persistent PD at follow-up (Paper II).
(1) Identify predictors of having PD after nine years among those with PD at the first examination.
2.5. Treatment and perceived treatment need of PD patients (Papers I and II).
(1) Describe the current PD treatment at follow-up after one year and after nine years and describe the perceived need for treatment after one year.
2.4. Effect of PD on chest pain, CAD and mortality (Papers I and III).
(1) Compare the PD and NoPD patients regarding persistent chest pain.
(2) Compare the death rate in the present sample to that of the general population.
(3) Compare PD and NoPD patients regarding cardiac events, risk factors for CAD and mortality.
2.6. Prevalence and outcome of nonfearful panic disorder (NFPD) in chest pain patients (Papers IV and V).
(1) Estimate the prevalence of NFPD at baseline and after nine years among the PD patients.
(2) Compare the patients with NFPD to those with PD with fear and those without PD regarding demographic characteristics, co-morbid psychiatric disorders, somatic disorders, chest pain, psychological distress, alexithymia, health care utilization and health related quality of life at baseline and after nine years.
Material and methods
3.1. Patient sample
3.1.1. Baseline
Between December 1994 and November 1996, 301 patients consecutively referred to one of four cardiology outpatient clinics in Oslo, Norway were asked to participate in an explorative study of psychological factors in chest pain patients. The patients had to meet the following inclusion criteria: 1) referred for investigation of chest pain as their main complaint;
2) no documented history of organic heart disease; 3) age 18–65 years; 4) no psychosis; 5) able to understand and write the Norwegian language; and 6) signed the informed consent form.
Documented organic heart disease included: a) previous myocardial infarction verified by ECG changes or enzyme examination, b) previous positive exercise test, thallium scintigraphy or coronary angiography, c) mitral valve prolapse verified by auscultation and echocardiography and d) arrhythmia documented by ECG.
Of the 264 patients who met the inclusion criteria, 199 agreed to participate in the study and these patients also constitute the baseline participants of the present study. There were no significant difference between participants (N=199) and non-participants (N=65) in terms of age, sex, coronary artery disease risk factors or prevalence of other medical disorder. The participants were diagnosed with coronary artery disease more often than the non-participants (16.1% vs. 3.1%; P=0.005). Of the 199 participants, 49 % were women and the mean age was 50,4 years (SD 9,4), 32 patients (16.1%) were diagnosed with coronary artery disease and 76 patients (38.2%) were diagnosed with panic disorder at baseline. Seven patients had both diagnoses (14).
3.1.2. One year follow-up
After one year, all study participants were asked to take part in a follow-up study. One patient had died and one patient had suffered a major stroke which made her ineligible. Of the 197 patients eligible for follow-up, 152 completed both the structured psychiatric interview and self-report questionnaires, 12 completed only self-report questionnaires and one completed the interview only. Thus, the participation rate was 77.2% (152/197) regarding complete data and 83.8% (165/197) regarding complete and incomplete data. Thirty-two patients did not participate in the follow-up study for the following reasons: Various disease states (n=4),
having moved to another part of the country or abroad (n=5), did not have time (n=1), or unknown reasons (n=22). We checked the Norwegian National Population Register for those who did not respond to any contact, but found no registered deaths.
3.1.3 Nine year follow-up
Eight to ten years after the baseline study, the participants were invited to a second follow-up. Fourteen patients had died and one had suffered a major stroke. Of the 184 eligible patients 150 participated in the follow-up study (82%). Of the 34 non-participants, seven had left the country, 12 could not be located, and 12 patients did not participate for the following reasons: Study not relevant to their condition at the time (n = 4), did not have time (n = 4), too difficult to come to the hospital (n = 1), afraid of giving away sensitive information (n = 1), disappointed with previous treatment at the hospital (n = 1), and unknown (n = 1). Three patients who filled in questionnaires did not attend psychiatric or cardiological evaluation sessions and were considered non-participants. Of the 150 participating patients, 12 attended the psychiatric evaluation session but not the cardiological evaluation session, which was scheduled to take place about a week after the psychiatric evaluation session. The reasons for their absence are unknown. They were still considered participants because information about their previous cardiac disorders could be obtained from their medical records.
3.2.0 Methods
3.2.1. Assessments of sampling bias
Using baseline data, the participating patients at the one-year and nine-year follow-up were compared to the non-participants to evaluate whether subject loss was systematic.
Regarding the one-year follow-up, the patients with complete data (n = 152) were compared to those who did not meet to the one-year follow-up or had incomplete data (n = 45).
The participating patients were significantly older than the non-participating patients (mean age 51.8 + 8.7 years vs. 46.2 + 10.5 years; P < 0.001). Otherwise, there were no significant differences between participants and non-participants regarding gender, prevalence of CAD, psychiatric disorders or scores on any outcome variable.
Regarding the nine-year follow-up, when assessing sampling bias among the eligible patients, there were no significant differences between the 150 participants and the 34 non- participants regarding sex, age, years of education or income at baseline. The participants did
not differ significantly from non-participants in prevalence of CAD (12% vs. 17%, respectively), PD (41% vs. 38%, respectively), or any of the outcome variables. We also compared the participants to the 15 non-eligible patients and found that the latter group were older mean age 50.1 + 8.9 years vs. 55.0 + 7.7 years; P=0.066) and had lower SF-36 scores (i.e. more HRQOL impairment) although it reached statistical significance regarding the physical functioning subscale only (mean 78.5+ 18.9 vs. 66.1+19.7; P=0.018). They had also significantly higher baseline SCL-90 somatization scores (mean 1.0 + 0.7 vs. mean 1.5 + 0.8;
P=0.033).
3.2.2. Procedure
The 197 patients eligible after one year were mailed a letter with an invitation to participate in a follow-up study, which comprised an appointment with a psychiatrist (TD), who had also conducted the baseline interviews, for the purpose of assessing psychiatric diagnosis and completing self-report questionnaires. The mean period between the cardiac outpatient investigation and the one-year follow-up was 12.3 months (range = 10.4–16.6, SD
=1.1).
From December 2003 to September 2005 the Norwegian National Population Register was searched for deaths among the baseline participants and the 184 eligible patients were mailed a letter with an invitation to participate in a long-term follow-up study. The invitation included details of an appointment with an experienced psychiatric resident trained in psychiatric interviewing (CBB), and an appointment for a cardiological evaluation at the outpatient clinic of the Department of Cardiology, Ullevål University Hospital. The mean time between the baseline investigation and the long-term follow-up was 8.6 years (range=8.1-9.9, SD = 0.35).
Data on non-fatal myocardial infarctions, percutaneous coronary interventions (PCI), and coronary artery bypass grafting (CABG) for the 12 participants who did not attend the cardiological evaluation and the 34 non-participating patients were obtained from their medical records. The causes of death of the 14 deceased patients were obtained from the National Death Register. The death rate in the study population was compared to the death rate calculated by Statistics Norway for a Norwegian population of equal age and gender distribution.
3.2.3. Psychiatric assessments
Psychiatric interview
Psychiatric disorders were assessed by using the Structured Clinical Interview for DSM IV- axis I disorders (SCID-I) (93) at baseline and at both follow-up investigations. The interview assesses current psychiatric state disorders (i.e. diagnostic criteria met within one month before interview) and lifetime psychiatric state disorders (i.e. diagnostic criteria met previously, but not last month). It was carried out by a psychiatrist, who is a trained interviewer (TD) at baseline and one year follow-up and by an experienced psychiatric resident (CBB) at the nine year follow-up. When a cardiological investigation was scheduled, the psychiatric interview was conducted before the cardiological appointment to ensure that the patients and the psychiatrist were blind to the results of the cardiologists. The psychiatrists were also blind to previous psychiatric diagnosis at the follow-up assessments. PD was recorded as current (diagnostic criteria met within one month before the interview) or in remission including PD in partial remission (diagnostic criteria met within the last 6 months prior to the interview, but not last month and only few symptoms of PD present) and PD in full remission (diagnostic criteria met previously, but not within the last six months and no symptoms of PD present). Major depression was recorded as current or lifetime, and the other diagnoses as current. The diagnoses were recorded immediately after the interviews. All interviews were audiotaped at both baseline and follow-up investigations. Thirty-five randomly selected tapes from baseline and twenty interviews from the one-year follow-up were assessed by an experienced
psychologist who was blind to the scores the interviewer had assigned. Thirty-two randomly selected tapes from the nine-year follow-up were assessed by TD who conducted the previous interviews. The interrater reliability scores were estimated for the diagnoses of panic disorder, major depression, generalized anxiety disorder, and somatoform pain disorder. The interrater reliability scores ranged from good to excellent for these psychiatric diagnoses at both baseline and follow-up (kappa 0.69–1.0). The kappa for panic disorder at the three evaluations was 0.88, 1.00 and 1.00 respectively.
Self-report measures
Demographic variables.
A self-report questionnaire comprising sex, age, marital status, education, work-status and income was filled in by all patients at baseline and reviewed by the investigator during the interview. At the follow-up interviews the patients were reassessed regarding marital and work status.
Psychological distress.
Symptom Checklist-90-Revised (SCL-90-R) (94) is a 90 item self-report questionnaire covering questions about a wide variety of psychological symptoms which are rated on a five point scale from 0 to 4 according to the level of experienced distress during the last week. The items generate nine dimensions: anxiety, phobic anxiety, depression, somatization,
obsessiveness, psychotisism, paranoididity, interpersonal sensitivity and hostility. The anxiety, depression and somatization subscales were considered most relevant and were therefore used in the baseline and follow-up investigations (14).
Suicidal ideation
Suicidal ideation was assessed by the SCL-90-R item 15: “During the last 7 days, how much were you distressed by thoughts of ending your life?” The responses were categorized as “0 = no thoughts” and all other responses as “1 = suicidal thoughts”).
Panic-Agoraphobia symptoms
Agoraphobic Cognitions Questionnaire (ACQ) (95) is a 20 item questionnaire which assesses beliefs about negative consequences of experiencing anxiety. The frequency of each belief is rated on a five point scale graded from 1=” The thought never occurs when I am anxious” to 5= “The thought always occurs when I am anxious”. ACQ was applied at baseline and at the long-term follow-up study.
Mobility Inventory for Agoraphobia (MIA) (96) contains 24 agoraphobic situations which are rated according to how often they are avoided on a five point scale from 1=”never avoided” to 5=”always avoided”. The situations are rated both as if they were encountered alone and as if they were encountered accompanied. MIA was applied at baseline and at the long-term follow- up study.
Hypochondriacal features.
Illness Attitude Scale (IAS) (97;98) is a 29 item self-report questionnaire rated on a five-point scale from 0 to 4 measuring health beliefs associated with hypochondriasis and attitudes
towards disease. Originally the scale comprised eight subscales. However, validations studies applying the scale to different medical patient samples have reported that the scale rather comprises two psychometrically sound subscales: Health Anxiety and Illness Behavior (95;99) which were used in the present study.
Symptom attribution
At baseline and after one year the patients were asked what they considered to be the cause of their chest pain. Cardiac disease, gastrointestinal disease, muscelo-skeletal disorder and psychological factors/stress were each rated on a global seven point scale from 1 = “not at all”
to 7 = “very high degree”.
Health related quality of life
The Medical Outcome Studies Short Form 36 (SF-36)(100) is a 36 item self-report questionnaire which generates eight subscales: physical functioning, role limitation due to reduced physical functioning, general health, vitality, body pain, social functioning, mental health and role limitation due to emotional problems. The subscales are calculated from the SF- 36 raw scores by use of a transformation manual which yield subscale scores with values from 0 to 100. A lower score indicate poorer quality of life. The scores decrease with increasing age, female gender and lower income (101). A psychometrically sound and validated Norwegian translation of the SF-36 was used in the present study (102).
Personality disorders.
The Personality Diagnostic Questionnaire for DSM-IV (PDQ-4) (103) is designed to assess ten personality disorders (paranoid, schizoid, schizotypal, histrionic, narcissistic, antisocial, borderline, avoidant, dependent and obsessive-compulsive) included in DSM-IV. The items assessing antisocial personality traits were omitted because some patients reported that they found the questions offensive. Thus, a 97 “true/false” item PDQ version was applied at baseline and used in the long-term follow-up study to assess predictors of still having PD after nine years among the baseline PD patients. At follow-up the following variables derived from PDQ were used: A total PDQ score and a diagnosis of the each of the personality disorders with a prevalence above 10% (avoidant (14%), depressive (14%), obsessive (27%) and paranoid (15%)).
Neuroticism
The neuroticism subscale of Eysenck Personality Questionnaire (EPQ)(104) was applied at baseline and used in the long term follow-up study to assess predictors of still having PD after nine years among PD patients. It comprises 23 “yes/no” items.
Alexithymia
Toronto Alexithymia Scale- 20 item version (TAS-20) (105) is a 20 item questionnaire that was applied at baseline to assess alexithymic features such as difficulties identifying and describing feelings and externally oriented thinking. It derives a TAS-20 total score and individuals scoring > 61 are considered alexithymic while individuals scoring < 51 are non- alexithymic. It was used in the first study of NFPD (paper IV)
Health care utilization and use of medication.
The use of medical facilities was assessed by the following questions: (1) “During the last 12 months have you contacted a health care provider because of chest pain? Yes/No”; (2) “If yes, indicate number of hospitalizations, emergency department visits, and visits at general practitioner, medical specialist, psychiatrist/psychologist and physiotherapist because of chest pain”; (3) “During the last 12 months have you contacted a health care provider because of other symptoms or diseases? Yes/No”; (4) “If yes, indicate number of hospitalizations, emergency department visits, and visits at general practitioner, medical specialist,
psychiatrist/psychologist and physiotherapist because of other symptoms or diseases”. The patients were also asked to list current use of medication.
Perceived treatment need.
Patients diagnosed with PD at the one year follow-up were asked if they received any current PD treatment, whether they experienced a need for treatment of PD symptoms, and if they wanted a report including information on PD, and general treatment recommendations to be sent to their GP.
Chest pain
At the follow-up investigations a self-designed chest-pain form was applied that included the following: (1) Chest pain remission (“Has your chest pain totally disappeared, meaning you have not had chest pain during the last four weeks? Yes/no”); (2) Experience of change in chest pain since baseline (0 = no chest pain, 1 = better, 2 = no change, 3 = worse); (3)
Perceived explanation of chest pain (“Did you get any explanation of chest pain? Yes/no”); (4) Degree to which patients’ perceived the explanation of chest pain to be sufficient (1 = not at all, 7 = very high degree). At the one-year follow-up study the form also included aspects of illness perceptions: (a) Perception of future (6 months) outcome of chest pain (0 = no pain, 7 = much worse); (b) What the patient can do to control the chest pain (1 = nothing, 6 = a lot); (c) What health personnel, such as doctors, can do to control the chest pain (1 = nothing, 6 = a lot).
Short Form McGill Pain Questionnaire (SF-MPQ) (106) contains 15 items that describe the quality of chest pain and was applied at baseline and both follow-up studies. Each chest pain quality is rated on a four point scale from 0=”no” to 3= “severe”. The questionnaire derives a total SF-MPQ score, a sensory pain component consisting of 11 items and an affective pain component consisting of four items. The intensity of chest pain was reported on a Visual Analogue Scale from 0-100.
3.2.4. Cardiological assessments
Nine year follow-up evaluations were conducted by four cardiologists, one of whom also took part in the baseline study.
Diagnostic assessments.
In all patients, a standard bicycle ergometer test was performed according to Nordenfelt et al. (107) at baseline and at the nine year follow-up. The cardiologist conducting the test interpreted the result. The test was considered coronary artery disease positive (CAD +) if ST segment depression t 1 mm occurred in any of the ECG leads during exercise. In addition, the appearance of typical chest pain, increasing ventricular ectopic beats, or absence of an increase t30 mm Hg in systolic blood pressure during exercise contributed to the diagnosis. When none of these signs was present, the test was considered coronary artery disease negative (CAD–). If inconclusive, the test was classified as such, and the patient was referred for further tests, such as thallium scintigraphy or coronary angiography. Referral for such investigations was determined by the cardiologist. The cardiologists were not informed of the results of the psychiatric interviews. For the purpose of final diagnostic classification, all chest pain records and results of cardiological investigations were reviewed by an independent cardiologist who
was blind to the results of the psychiatric interviews. This review did not result in any changes to the initial classification of the diagnoses.
Registration of medical diseases, medication, coronary artery disease risk factors and chest pain characteristics.
A cardiological assessment form was completed by the cardiologist for each patient at the baseline and follow-up examinations. The form recorded data on the patients’ previous or prevailing medical diseases, medication and risk factors for CAD (family history of CAD before age 65 years in parents or siblings, smoking habits, diabetes, hypertension, obesity [body mass index > 26 kg/m2] or dyslipidaemia [total cholesterol > 6.5 mmol/L and/or triglycerides > 2 mmol/L]). Chest pain was recorded with regard to presence/absence, location, pain characteristics, effect of nitroglycerine, relief by rest, chest wall tenderness and duration, and the chest pain was assigned as typical or atypical for CAD.
3.2.5. Statistical analysis
The SPSS/PC statistical package was used for all data analyses (version 10.0-14.0). For multiple comparisons a significance level of 1% was applied, otherwise the significance level was 5%.
Comparisons between groups of patients were performed using the chi-square test for dichotomous variables. The independent sample Student’s t test was used for normally distributed continuous variables and the Mann–Whitney U test was used for continuous variables without normal distributions. The latter variables were also analyzed using Student’s t test. As the two tests yielded nearly identical results, data are presented as the mean ± SD. All tests were two tailed. Differences between three groups were analyzed using One-Way
Analyses of Variance (One-Way ANOVA). Scheffes test was applied for multiple comparisons.
ThePaired-Sample T test assessed change in scores of continuous variables separately for each study group.
Agreement between interviewers on psychiatric diagnoses was assessed using the kappa coefficient. To control for age, gender and education regarding the SF-36 scores linear regression analyses were conducted with SF-36 scores as the dependent variable. Baseline variables that were suspected to be associated with current PD at follow-up were entered one at a time in a binary logistic regression analysis with current PD at follow-up as the dependent variable. Only the variables that were statistically significantly related to PD at follow-up (P <
0.01) were considered for further analyses for model development. Variables with high intercorrelations (r >0.7) were excluded. A cut-off value of the predictors was established at a level of optimal sensitivity and specificity by inspection of the ROC curve.
Statistics Norway estimated the mortality rate of the entire Norwegian population during the period in which the study was conducted. The standardized mortality ratio (SMR) (observed deaths ÷ estimated deaths) was estimated with a 95% confidence interval (CI). The 95% CI of the SMR was calculated from the following equation: (e–1,96/ D(u) * SMRu, e1,96/ D(u) * SMRu) where D(u) is the observed number of deaths in the study population.
3.2.6 Ethical aspects.
The research protocol including the follow-up study was accepted by the Regional Ethics Committee, Oslo, in November 1994. Both at baseline and follow-up, all patients signed a consent form where they also agreed to being contacted again for future follow-up
participation. On the consent form it was explicitly stated that participation in the study was voluntary, and the patients had the rights to see the data registered on them and have the data removed on any occasion. The patients received written information about the purpose of the study before all three investigations. In this information it was also emphasized that study participation was voluntary and that non-participation would have no consequence for future contact with the Cardiology outpatient clinic.
As the SCID interview contains personal questions that may be distressful for some to answer, the investigators tried to be sensitive to the patients responses. The interview of a previously traumatized patient was thus interrupted because the questions evoked unpleasant memories. Otherwise, none of the patients reported that they felt uncomfortable during the interview. In five cases the patients were referred by the investigator to treatment in a
psychiatric outpatient clinic because severe psychiatric disorder was uncovered, three of which reported recent suicidal ideation. When somatic disorders requiring further medical
investigation or treatment were diagnosed during the cardiological investigation, the
cardiologist referred the patient to such treatment or informed the general practitioner about the necessity of further investigation in a written report.
4.0. Summary of papers.
Paper I
A 1-year follow-up study of chest-pain patients with and without panic disorder.
Gen Hosp Psychiatry 2006 Nov;28(6):516-24.
The first paper is a follow-up of 199 chest pain patients assessed at a cardiology outpatient clinic one year previously. At the baseline investigation 32 patients (16%) suffered from coronary artery disease (CAD) and 76 patients (38%) suffered from panic disorder (PD).
The aims of this study were to: (1) study the persistence of PD after one year; (2) investigate the association between PD at baseline and outcome in terms of chest pain,
psychiatric morbidity, psychological distress, suicidal thoughts and health related quality of life (HRQOL); (3) study the course of pain, distress, symptom attribution and HRQOL; and (4) describe treatment and perceived treatment needs of PD patients.
The follow-up assessments included a psychiatric interview (SCID I) and self-report questionnaires (McGills Pain Questionnaire (MPQ), Symptom Checklist-90-R (SCL-90-R), Illness Attitude Scale (IAS), and The Medical Outcome Study Short Form-36 (SF-36)). After one year, one patient had died and one patient had suffered a major stroke. Of the 197 eligible patients, 153 (78%) participated in the follow-up investigation.
Among the participants were 55 patients diagnosed with PD at baseline and 43 of these (78%) still suffered from PD at follow-up. PD at baseline was associated with chest pain persistence and psychiatric morbidity (current major depression, pain disorder and simple phobia) at follow-up. The baseline PD patients had also significantly higher follow-up scores of SCL-90-R anxiety, depression and somatization, hypochondriasis (IAS), and lower scores on seven of the eight SF-36 dimensions of HRQOL compared to NoPD patients. Still, after one year only 6% of the PD patients used effective treatment and 3% reported a treatment need at follow-up.
The results suggest that only a minority of chest pain patients with PD receive anti-panic treatment one year after they are diagnosed with PD and despite continual panic attacks, chronic distress and impairment.
Paper II
Nine-Year Follow-Up of Panic Disorder in Chest Pain Patients: Clinical Course and Predictors of Outcome. Psychosomatic Medicine. Submitted.
The aims of this study were to: (1) study the persistence of panic disorder (PD) after nine years; (2) investigate the association between PD and long-term outcomes in terms of
psychiatric morbidity, psychological distress, and health-related quality of life (HRQOL); (3) identify predictors of having persistent PD after nine years among baseline PD patients.
The nine-year follow-up assessments included a psychiatric interview (SCID I) and the same self-report questionnaires as included at the one-year follow-up. After nine years 14 patients were deceased and one had suffered a major stroke. Of 184 eligible patients, 150 (82%) participated in the follow-up study.
Among participants with PD at baseline (N = 55), 14 (25.5%) suffered from persistent PD at follow-up. PD at baseline was associated with a higher prevalence at follow-up of co- morbid axis I disorders (hypochondriasis and major depression), higher psychological distress, and poorer HRQOL compared with patients without PD at baseline. A mean baseline SCL-90- R somatization score above 1.4 predicted a fivefold increased risk of having persistent PD at follow-up. Patients with persistent PD had particularly poor outcomes regarding co-morbid axis I disorders, suicidal ideation (21%), psychological distress, and HRQOL. After nine years, 31% of the patients with persistent PD and 15% of patients with PD in remission received pharmacological antipanic treatment and none received cognitive therapy.
The results suggest that despite a high remission rate of PD in patients with chest pain, many PD patients still suffer from significant psychiatric morbidity and quality of life
impairment after nine years. PD patients with high somatization scores require special attention because they have an increased risk of having continual panic attacks and subsequently particularly poor outcomes.
Paper III
A long-term follow-up study of chest pain patients: Effect of panic disorder on mortality, morbidity, and quality of life. Cardiology. In press.
The aims of the third paper were to (1) compare the mortality rate of chest pain patients with that of the general population; (2) examine the long-term relationship between PD at baseline and mortality, CAD, and coronary risk factors at follow-up; and (3) compare the
outcomes of chest pain patients with or without PD at baseline in terms of chest pain, anxiety, depression, and HRQOL nine years after the initial baseline examination.
One hundred and ninety-nine patients consecutively referred to a cardiological outpatient clinic because of chest pain were reassessed after nine years by a cardiological examination including a bicycle exercise test and recording of risk factors for CAD. The patients also filled in self-report questionnaires regarding chest pain, anxiety and depression (SCL-90-R) and HRQOL (SF-36). Statistics Norway compared the death rate in the sample to that of the Norwegian general population. At the initial examination 32 patients (16%) suffered from CAD and 76 patients (38%) from PD.
The death rate in the study population was not significantly different from that in the general population and no significant associations were found between PD at baseline and mortality and cardiac morbidity at follow-up. PD at baseline was associated with significantly higher follow-up scores of chest pain intensity, depression and anxiety; and lower SF-36 scores indicating HRQOL impairment (physical functioning, role physical, body pain, and general health).
The length of follow-up may have been too short to find an increased death rate in this patient sample; however, the results suggest that PD has a negative long-term effect on psychological and physical well-being of chest pain patients which emphasize the necessity of identifying PD patients and offering them adequate treatment.
Paper IV
Nonfearful panic disorder in chest pain patients. Psychosomatics 2004 Jan;45(1):69-79.
The aims of this study were to: (1) Investigate the prevalence of nonfearful panic disorder (NFPD) among chest pain patients with PD; and (2) Compare patients with NFPD to PD patients reporting fear and NoPD patients regarding demographic variables, self-reported anxiety, agoraphobia and somatization, co-morbid axis I disorders including somatoform disorder, presence of somatic disorders, alexithymia and health related quality of life in 199 chest pain patients consecutively referred to cardiological outpatient investigation.
Seventeen (22%) of the PD patients fulfilled the criteria for NFPD. Demographically no significant difference was found between the patients with NFPD and those with PD with fear.
The NFPD patients differed from PD patients in lower scores on self-reported panic symptoms and lower frequency of agoraphobia. Presence of somatoform disorder, alexithymia or somatic disorder did not explain NFPD symptoms; yet, the prevalence of somatic disorders was high in
this group. Our findings confirm that NFPD is a subgroup of PD which needs to be identified as their health related quality of life impairment is not significantly different from that of PD patients reporting fear.
Paper V
Nonfearful panic disorder in chest pain patients: status after nine-year follow-up.
Psychosomatics. In press.
The fifth paper is a nine year follow-up study of the patients described in paper IV. The aims of this study were to: (1) Explore the concept of NFPD by making diagnostic
reassessments after nine years, and (2) Investigate the long-term outcome of NFPD patients in terms of psychiatric co-morbidity, psychological distress, chest pain, somatic diseases, health care utilization and HRQOL.
We studied 199 patients previously referred to cardiology outpatient investigation because of chest pain. At baseline a total of 76 patients suffered from PD of which 17 fulfilled criteria for NFPD. Eleven patients with baseline NFPD and 44 patients with PD with fear at baseline met to the follow-up investigation. The follow-up assessments included a
cardiological and psychiatric investigation.
After nine years, no patients suffered from NFPD but seven (64%) of the patients with baseline NFPD fulfilled diagnostic criteria for either current or lifetime PD with fear. There was no significant difference between the baseline NFPD and PD patients regarding prevalence of psychiatric co-morbidity, or self-report of chest pain, health care utilization, and HRQOL at follow-up. These results strongly suggest that NFPD is a subgroup of PD which needs to be recognized as such to initiate treatment and avoid long-term impairment. Future studies should investigate the effect of recommended PD treatment on NFPD patients.
General discussion
5.1.0. Methodological issues.
5.1.1. Sample
The study is the most extensive follow-up study of chest pain patients with panic disorder to date. The sample size is larger than in the study by Beitman and colleagues (15) who followed 72 chest pain patients of whom 36 suffered from PD; and the participation rate is considerably higher than in the study by Fleet and colleagues (16) where only 54% took part in the follow-up assessments. The participation rates in the present study were 77% (153/199) at the one year assessments and 75% (150/199) at the nine year assessments regarding the total sample and 78% (153/197) and 82% (150/184) respectively regarding eligible patients (deceased and stroke patients excluded). We considered the size of the sample at follow-up large enough to determine the prevalence of persistent PD among the patients with PD at baseline and to compare the PD and NoPD patients regarding outcome measures. However, regarding prevalence of cardiac events and deaths in the PD and NoPD patients, the sample was considered too small to make statistical comparisons between the study groups (paper III).
The sample size was also considered insufficient to make separate comparisons of the death rate in each of the study groups to that of the general population (paper III).
Generalizability.
The question about generalizability of the results concerns both the loss of patients from baseline to follow-up (internal validity) and to what extent the baseline sample is representative for patients with chest pain that present at the cardiology outpatient clinic. Another issue is to what degree the results may be generalized to other groups of chest pain or PD patients (external validity).
Internal validity
A methodological issue encountered in follow-up studies is whether the participating patients are representative for the initial sample studied or whether a systematic subject loss has occurred. Theoretically one might assume that patients with more somatic symptoms, more worry about their health, and more disease conviction or true medical disorders would be more
